This invention relates generally to semiconductor device fabrication, and more particularly to the use of phase shift masks (PSM""s) in conjunction with such fabrication.
Since the invention of the integrated circuit (IC), semiconductor chip features have become exponentially smaller and the number of transistors per device exponentially larger. Advanced IC""s with hundreds of millions of transistors at feature sizes of 0.25 micron, 0.18 micron, and less are becoming routine. Improvement in overlay tolerances in photolithography, and the introduction of new light sources with progressively shorter wavelengths, have allowed optical steppers to significantly reduce the resolution limit for semiconductor fabrication far beyond one micron. To continue to make chip features smaller, and increase the transistor density of semiconductor devices, IC""s have begun to be manufactured that have features smaller than the lithographic wavelength.
Sub-wavelength lithography, however, places large burdens on lithographic processes. Resolution of anything smaller than a wavelength is generally quite difficult. Pattern fidelity can deteriorate in sub-wavelength lithography. The resulting features may deviate significantly in size and shape from the ideal pattern drawn by the circuit designer. For example, as two mask patterns get closer together, diffraction problems occur. At some point, the normal diffraction of the exposure rays start touching, leaving the patterns unresolved in the resist. The blending of the two diffraction patterns into one results from all the rays being in the same phase. Phase is a term that relates to the relative positions of a wave""s peaks and valleys. In FIG. 1A, the waves 102 and 104 are in phase, whereas in FIG. 1B, the waves 106 and 108 are out of phase.
One way to prevent the diffraction patterns from affecting two adjacent mask patterns is to cover one of the openings with a transparent layer that shifts one of the sets of exposing rays out of phase, which in turn nulls the blending. This is shown in FIGS. 2A and 2B. In FIG. 2A, the mask 202 causes an undesirable light intensity as indicated by the line 204. In FIG. 2B, adding the phase shifter 206 to the mask 202 causes a desirable light intensity as indicated by the line 208. This mask 202 in FIG. 2B with the phase shifter 206 added is a phase shift mask (PSM), which is a special type of photomask.
A typical photomask affects only one of the properties of light, the intensity. Where there is chromium, which is an opaque region, an intensity of zero percent results, whereas where the chromium has been removed, such that there is a clear or transparent region, an intensity of substantially 100 percent results. By comparison, a PSM not only changes the intensity of the light passing through, but its phase as well. By changing the phase of the light by 180 degrees in some areas, the PSM takes advantage of how the original light wave adds to the 180-degree wave to produce zero intensity as a result of destructive interference.
PSM""s have gained increased popularity among manufacturers as the feature sizes they are tasked with printing become smaller, and the topography over which these features must be printed becomes more varied. PSM""s offer their customers the opportunity to greatly improve the resolution capability of their steppers. This allows them to print smaller feature sizes using the same equipment and processes.
One particular type of PSM is referred to as an alternating PSM. The PSM of FIG. 2B was one example of an alternating PSM. In an alternating PSM, closely spaced apertures are processed so that light passing through any particular aperture is 180 degrees out of phase from the light passing through adjacent apertures. Any light that spills over into the dark region from the two edges that are out of phase will destructively interfere. This reduces the unwanted exposure in the center dark region.
FIG. 3 shows another example of an alternating PSM, and more specifically, a dual-trench alternating PSM 300. The PSM 300 includes two layers, a chromium layer 302, and a quartz layer 304. The chromium layer 302 is the same type of layer typically found in other, non-PSM photomasks, in which light is exposed therethrough to an underlying semiconductor wafer. Clear regions within the chromium layer 302 allow light to pass through, whereas opaque regions within the chromium layer 302 prevent light from passing through. The clear and opaque regions are arranged to correspond to a desired semiconductor design, or pattern. In the PSM 300, there are clear regions 306A, 306B, 306C, 306D, and 306E.
The quartz layer 304 is more generally a clear or transparent layer, in which different-sized trenches are alternatively added beneath the clear regions of the chromium layer 302 to phase shift light passing through these clear regions. For instance, the alternating clear regions 306A, 306C, and 306E of the chromium layer 302 have shallow trenches beneath them in the quartz layer 304. Conversely, the alternating regions 306B and 306D of the chromium layer 302 have deep trenches beneath them in the quartz layer 304. The PSM 300 is an alternating PSM in that the PSM design repeats on an alternating basis for clear regions of the chromium layer 302, such that one clear region has a shallow trench beneath it, whereas the next clear region has a deep trench beneath it, and so on. The PSM 300 is a dual-trench alternating PSM in that there are two trenches that repeat, a shallow trench and a deep trench.
The manner by which the PSM 300 of FIG. 3 can be fabricated according to the prior art is summarized with reference to FIGS. 4A-4I. In FIG. 4A, the clear regions within the chromium layer 302 are already present, by a process of photoresist patterning in which the photoresist is first exposed by e-beam writing and then developed, etching the chromium layer 302, and then stripping the remaining photoresist. A new layer of photoresist 402 has been added, such as by a coating process. In FIG. 4B, the layer of photoresist 402 is exposed to correspond to the deep trenches 306B and 306D. The exposure is accomplished by e-beam writing. In FIG. 4C, exposed areas of the photoresist 402 are developed to remove them, and in FIG. 4D, the quartz layer 304 is dry etched to initially form the deep trenches 306B and 306D.
The quartz layer 304 is dry etched for 60 degrees, so that if there are defects within the quartz layer 304, only 60 degrees of such defects will be present. This amount is minimal, and will not affect printing of the semiconductor device using the PSM 300. The process of FIGS. 4A-D is repeated for a total of three times, so that a total of 180 degrees of phase shift is achieved within the quartz layer 304.
In FIG. 4E, another layer of photoresist 402 is added, such as by a coating process. In FIG. 4F, the layer of photoresist 402 is exposed to correspond to all the trenches 306A, 306B, 306C, 306D, and 306E. The exposure is accomplished by e-beam writing. In FIG. 4G, exposed areas of the photoresist 402 are developed to remove them, and in FIG. 4H, the quartz layer 304 is dry etched to completely form the deep trenches 306B and 306D that were previously initially formed, as well as to completely form the shallow trenches 306A, 306C, and 306E.
The quartz layer 304 is again dry etched for 60 degrees, so that if there are defects within the quartz layer 304, only 60 degrees of such defects will be present. The process of FIGS. 4E-4H is repeated for a total of four times. Once the fourth time is finished, the remaining photoresist 402 is removed, such that the final PSM 300 remains, as has already been shown in FIG. 3.
This conventional approach to manufacturing the dual-trench alternating PSM 300 has several disadvantages, however. Overlay errors can error between successive exposures of the photoresist, such as between successive e-beam writings. These overlay errors induce anti-reflective layer loss around the etched regions, which can be difficult to discover during inspection. Because photoresist is exposed for a total of eight times, there are eight such successive e-beam writings, and the potential for overlay error inducing anti-reflective layer loss is great. Furthermore, e-beam writing in particular can be a time-consuming and/or costly process, such that fabricating the PSM 300 will be a lengthy and/or costly process since e-beam writing is used eight times.
Therefore, there is a need for a process for fabricating a dual-trench alternating PSM that overcomes the problems associated with manufacturing such PSM""s in the prior art. Such a process should minimize the use of e-beam writers as much as possible. Minimizing the use of e-beam writers will allow such a process to avoid as much as possible overlay errors inducing anti-reflective layer loss. Furthermore, such minimization will decrease the time and/or the cost in fabricating dual-trench alternating PSM""s. For these and other reasons, there is a need for the present invention.
The invention relates to fabricating a dual-trench alternating phase shift mask (PSM). An opaque layer over a transparent layer of the PSM is patterned according to a semiconductor design. The opaque layer may be a chromium layer, whereas the transparent layer may be a quartz layer. The transparent layer is dry etched a first number of times through a first photoresist layer applied over the opaque layer and patterned according to the deep trenches of the alternating PSM design by using beam writing. This initially forms deep trenches of the PSM. The transparent layer is dry etched a second number of times through a second photoresist layer applied over the opaque layer and patterned according to the deep trenches and the shallow trenches of the alternating PSM design by using backside ultraviolet exposure. This completely forms shallow trenches and the deep trenches of the PSM. The second photoresist layer is then removed.
The invention provides for advantages not found within the prior art. E-beam writing is significantly reduced in fabricating a dual-trench alternating PSM, as compared to the prior art, by instead using backside ultraviolet exposure. This significantly decreased overlay error, which means that induced anti-reflective layer loss is also significantly reduced. The reduction in use of e-beam writing also means that the invention generates a dual-trench alternating PSM in a less costly and/or less time-consuming manner as compared to the prior art. Other advantages, embodiments, and aspects of the invention will become apparent by reading the detailed description that follows, and by referencing the attached drawings.